2058 JACK L. RYAN Inorganic Chemislvy

The values of Xuncor(M) obtained from the data of Table VI using eq 2 are listed in Table VI1 along with those obtained for the MClf13- and MBrfi3- coiiiplexes from ref 4. It is interesting to note that there is a

x , ~ ~ ~ ~ ~ ( M ) for a given lanthanide as one goes from 1968.

hexachlorides through hexabromides to hexaiodides. A similar effect of about the same degree occurs for tetravalent actinide hexahalides and uranyl tetraha- Mess and for uranium(Vi) hexahalides.'*

slight increase in all cases in the measured value of (18) J. L. Ryan, unpublished data presented in part a t the 155th National Meeting of the American Chemical Society, San Francisco, Calif., April 1-.5,

CONTRIBUTION FROM PACIFIC NORTHWEST LABORAI.ORICS, 99352 ~ A T T E L L E MEMORIAL INSTITUTE, RICHLAND, WASHINGTON

Weak or Unstable Iodo Complexes. 11. Iodo Complexes of Titanium( IV), Iron( I11 ), and Gold (111)'

BY JACK L. RYX?:

Received March 3, 1969

Salts of iodo complexes of several d group metal ions which are strongly oxidizing toward the iodide ion were prepared by reaction of the corresponding chloro complexes with liquid anhydrous hydrogen iodide. These unstable complexes are TiIe2-, FeId-, and AuId-. An attempt to prepare CUI**- by this method was unsuccessful. Of these, only FeI4- has any stability at all in solution at 25", and i t decomposes with a half-life of a few minutes. The absorption spectra of these complexes are compared with the spectra of the corresponding chloro and bromo complexes. Evidence is presented for the lack of higher halide complexes of Au( 111) than AuXd-.

Introduction The preceding paper2 discusses the preparation of

very weak iodo complexes of the trivalent lanthanides by condensing liquid anhydrous H I on the correspond- ing salts of the chloro complexes. The driving force for the reaction is the higher free energy of formation of HC1 than of HI and to a lesser extent the greater volatility of HC1. Partly because of the lorn tempera- tures involved, this technique is also very useful in the preparation of relatively stable salts of halide com- plexes where the reducing power of the halide is such that the particular metal valence state cannot be maintained in solution in the presence of the halide. Thus salts of CeBrs2- have been prepared by a similar r n e t h ~ d . ~ , ~ This paper discusses the preparation and absorption spectra of salts of the TiIf12- complex which can be expected to be both very weak and also unstable to autoreduction of the central metal and of the Fe14- and A d 4 - complexes which are expected to be unstable to autoreduction. It appears from preliminary work that it will be possible to prepare salts of chloro complexes of very strongly oxidizing metal ions by treat- ing the fluor0 complexes with Sic14 or BC13 in liquid HC1 a t low temperatures.

Experimental Section Salts of the iodo complexes were prepared and the absorption

spectra of these solid salts obtained by the methods used for the lanthanide complexes.2 The absorption spectrum of (C2Hs)a- NFeIa was also obtained in solution in CaSOa-dried acetone and nitromethane. Molar extinction coefficients of FeIa- were ob- tained by dissolving a weighed sample of the salt in a known vol-

(1) This paper is based on work performed under Contract Nu. AT- (45-1)-1830 for the U. S. Atomic Energy Commission.

(2) J. L. Ryan, Iizorg. Chem., 8 , 2053 (1969). (3) J. L, Ryan and C. K. J@rgensea, 3. Phys . Chem., 70, 2815 (1966).

ume of acetone a t -78". This solution was then added to a 0.1- cm cell and the spectrum was scanned rapidly immediately after the transfer. Small corrections were made for decomposition based on the rate observed at a constant wavelength on another sample of the solution. ,411 spectra were obtained using a Cary Model 14 recording spectrophotometer.

Preparation of Complexes Ti(1V) is considered to be one of the most typical

A group acceptors4 and thus iodo complexes, if such exist, are expected to be very weak. The absorption spectra (due to electron transfer) of TiClO2- and TiBrfi2- have been recently m e a s ~ r e d . ~ ~ ~ From these and the variation of optical electronegativity of the halide,7 i t can be predicted that Ti162- mill be black and relatively unstable to decomposition to Ti(II1) and iodine. When [(C2H6)4N]2TiC16 is treated with liquid HI, a pure black solid resembling carbon black results. This material is very unstable to air and is changed to the dark red-brown of Is- in the time required to pour it quickly from one container to another in laboratory air. In this regard i t is less stable than the lanthanide MIo3- salts.2 It is not stable in solvents such as acetonitrile or nitromethane a t 25" with or without excess I-. If anhydrous H I is added to a partially frozen (about -46') acetonitrile solution of (CIHs)4NI containing a small amount of [(C2Hb)4N]2TiC16, an intense black solution results. The intense black color is completely replaced in a few seconds by the much less intense reddish brown color of Is-. (In the absence of the Ti salt only a white precipitate similar

(4) S. Ahrland, Strztct. Buiadirig (Berlin), 1, 207 (1966). ( 5 ) J. I,. Ryan, unpublished results quoted by C. K. JZrgeneen. (ti) B. J. Brisdou, T, 13. Lester, aud I< . A . \\'nlton, Specli .uchiin. A c / < r , ZSA,

(7) C. K. J<ii-gmseti, "Jtiorgariic C'otnl,lexrr," Acacleniir Pres?, I.oud~m, I969 (1967).

1963.

Vol. 8, No. 10, October 1969 WEAK OR UNSTABLE I O U 0 COMPLEXES 2059

to that formed with HBr3 forms when H I is added a t the freezing point of the solvent, but a t 25” some 1 3 -

forms due to H I reaction with the solvent.) It thus appears that Ti162- is a sufficiently strong complex to be prepared in an acetonitrile solution containing excess iodide and anhydrous HI but that it is very un- stable toward reduction of the Ti(1V). The T i I P complex then appears to be stronger than the lanthanide MIe3- complexes2 which are not obtained a t all under these conditions.

It should be noted that a pyridinium salt of TiIe2- has been previously reported as a “dark brown” solid isolated from concentrated aqueous When C5H5- NHI in concentrated H I is added to the dark reddish brown solution obtained by dissolving hydrated Ti02 in concentrated HI, a dark brown solid is isolated having the reported properties. It is not appreciably soluble in water but is soluble in acetone indicating that it is CsH5NHI3. When black [ ( C ~ H E ) ~ N ] ~ T ~ I ~ is treated with cold, 12-free, aqueous, concentrated HI, it is immediately converted to a dark brown solid which again is not water soluble but is acetone soluble and apparently is (C2Hj)lN13. From this and the ob- served air sensitivity of the [ (C2Hb)4N]2Ti16 prepared in this manner, there is little doubt that the previous reports of a TiIG2- salt is in error. It appears that Ti(1V) is not stable in concentrated aqueous HI.

In order to demonstrate further the utility of this method for the preparation of iodo complexes of moderately strongly oxidizing metal ions, the iron(II1) system was chosen. Pure Fe(II1) iodide compounds are unknown and i t has been stated that “iron(II1) is too strong an oxidizing agent to coexist with a good reducing agent like I-.”9 Although Fe(II1) is con- sidered to be an A group acceptor,4 its chemistry (stable complexes with thiocyanate, etc.) indicates that it is certainly not as strongly A group in character as trivalent lanthanides and Ti(1V). Because of this and since FeBr4- salts are readily prepared even from aqueous HBr solutions, there is no reason to believe that Fer4- would not be a moderately strong complex. Thus there is every reason to believe that barring oxidation state stability problems, if the chlorides of an FeCI4- salt were replaced by iodides, the resulting material would be a salt of FeI4-. When (CzH5)4- NFeC14 is treated with liquid anhydrous HBr, it is com- pletely converted to (CzH6)4NFeBr4 as determined by the identity of the absorption spectrum of the product to that of (C2H5)4NFeBr4 prepared from aqueous HBr.

The electron-transfer spectra of FeC14- and FeBr4- have been published.’O From these i t can be predicted that the lowest molar extinction coefficient for Fe14- in the visible region would be about 4000 and thus salts of FeI4- would be black. When (C2H6)4NFeC14 is treated with anhydrous liquid HI, a black crystalline solid is obtained. This material is considerably more

(8) H. J. Emeleus and G. S. Rao, J. Chem. Soc., 4245 (1958). (9) F. A. Cotton and G. Wilkinson, “Advanced Inorganic Chemistry,”

2nd ed, Interscience Division, John Wiley & Sons, Inc., New York, N. Y. , 1966, p 857.

(10) P. Day and C. K. Jylrgensen, J . Chem. SOC., 6226 (1964).

stable than the Ti(1V) or lanthanide(II1) hexaiodo complexes and appears to be stable in dry air. It can be handled for very short periods in laboratory air but is decomposed completely in a few hours by atmo- spheric moisture. It can be dissolved in dry acetone, acetonitrile, or nitromethane to produce an intensely purplish black solution. At 25’ the Fe(II1) is reduced with a half-life of a few minutes in these solvents, and this rate is greatly accelerated by excess I-. The decomposition rate is decreased markedly by decrease in temperature and the half-life is a t least several hours in acetone a t - 78”. Similar purplish black solutions of FeI4- can be obtained by adding anhydrous H I directly to acetonitrile solutions of (C2HJ4NFeC14 a t the freezing point of the solvent or to acetone solutions a t -78”. The color of these solutions is not affected by addition of excess I- as (C4H9)4NI.

Au(II1) is a very typical Chatt-Ahrland B group acceptor4 and as such should form very strong iodo complexes. Several 19th century references to gold(II1) iodide and iodoaurates exist. l1 All of these involve evaporation of aqueous iodide solutions thought to contain Au(II1). It is now known that Au(II1) is not stable in aqueous iodide solution^,'^-^^ and in fact the quantitative liberation of iodine by Au(II1) in aqueous iodide solutions has been used as an analytical proce- dure for gold.I4 The products obtained in these 19th century references probably consisted of mixtures of gold (I) iodide or complex iodides and salts of I 3 + .

When ( C ~ H ~ ) ~ N A U C ~ ~ (composition checked by analy- sis) is treated with liquid anhydrous HI, a crystalline black solid (resembling graphite) is obtained. Analysis for total gold content indicates complete conversion to ( C ~ H ~ ) ~ N A U I ~ , and qualitative turbidimetric C1- analy- sis indicated only a trace of residual C1-. The salt is moderately stable to laboratory air apparently because of its extreme water insolubility. This insolubility is related to the expected strength of the A d 4 - complex and its lack of tendency to hydrolyze (B group be- havior) and to the inherent low solubility of quaternary ammoaium salts of AuX4- complexes (see also the extremely high distribution coefficient of AuC14- onto quaternary ammonium anion exchanger^'^). The salt dissolves with immediate decomposition in nonaqueous solvents such as acetone, acetonitrile, and nitromethane. At -78” the salt dissolves t o a very slight extent in acetone and to a greater extent in propylene carbonate to produce violet solutions which change to brown in a few minutes indicating that A d - is very unstable in solution. The solid is also violet when streaked on porcelain.

Several insoluble “double auric iodides” have been reported as having been prepared from aqueous solution

(11) J. W. Mellos, “A Comprehensive Treatise on Inorganic and Theo- retical Chemistry,” Vol. 111, Longmans, Green and Co., New York, N. Y.,

(12) A. K. Gangopadhayay and A. Chakrovorty, J . Chem. Phys., 36, 2207 (1961).

(13) Reference 9, p 1048. (14) W. M. Latimer and J. H. Hildebrand, “Reference Book of Inorganic

(15) K. A. K r w s and F. Nelson, PYOC. Itzteun. Coizf. Peaceful Uses At.

1946, pp 609-610.

Chemistry,” 3rd ed, The Macmillan Co., New York, N. Y., 1951, p 127.

Energy, Geneva, 1955, 7 , 113 (1966).

2080 JACK L. RYAN Inorgunic Chemistry

TABLE I ELECTROS-TRANSFER SPECTRA OF TiX& a

TiC12- (25.0) [-1300], (29.6) [-17,000], 31.8 [22,000], 43 8 [30,000] TiBrh2- (17.7) [-700], (21.7) [ -SOOO], 24.9 [16,000], (27) [ ~ 1 2 , 0 0 0 ] , 36 .1 [20,000] [ ( C ~ H ~ ~ N I Z T ~ I I (9.5) [ O . 31, 12.1 [0.65], 14 .3 [0.651, (18.7) [ 0 . 8 ] , 2 3 . 2 [1.0]

a Wave numbers in kilokaisers with shoulders in parentheses. For TiCls2- and TiBr& values in brackets are molar extinction coefficients and for [ ( C ~ H ~ ) ~ N ] ~ T ~ I O values are intensities relative to the strongest listed transition taken equal to 1.0.

by adding auric chloride to the appropriate iodide solutions. If halogen exchange and precipitation are fast enough relative to reduction of Au(II1) and the precipitate is sufficiently insoluble, this method might work. It should be noted though that the same cations which give insoluble salts with A d - can be expected to give illsoluble Is- salts and the resulting mixture of insoluble AuI and MI3 will appear analytically to be a salt of A d - . The reported preparation of C S A U I ~ ~ ~ was repeated and the black precipitate first formed was found to begin to turn brown immediately and this continued during filtering and drying. A similar prep- aration was carried out using triethylmethylammonium iodide in which case the solubility and formation of 13- appeared to be much less. Three preparations all con- tained Au(II1) iodo complexes as determined by the absorption spectra. The spectra varied and only one was close to that of (CzH6)4NAu14 prepared as discussed above. These preparations from aqueous solution all contained more residual chloride than the (C2H5)4N4uI4 prepared by reaction of the chloro complex with HI . This is not surprising in view of the fact that such salts of both AuC14- and A d 4 - have very low solubility, and jn this aqueous method iodide will be removed from the mixing zone more rapidly than the quaternary ammo- nium ion. Thus the aqueous method cannot be ex- pected to yield pure salts of Xu1,- but instead will give mixed chloro-iodo complexes generally contaminated with Is- and AuI or salts of Au12-.

An attempt was made to prepare a salt of the hypo- thetical ion CuId2- by treating a salt of CuCla2- with liquid HI. Salts of this complex, if such exist, would be black. When HI was condensed onto [(CZHJ4N]Z- CuC14, the color immediately became that of 1 3 - in- dicating reduction of the Cu(I1).

Absorpticjn Spectra The absorption spectra (due to electron transfer) of

TiC16*- and TiBra2- have been measured by this author as well as by Brisdon, et a1.,6 who along with Jgr- gensen,17'ls using the data of ref 5, have discussed the spectra in detail. Band assignments have been pro- posed by J$Jrgensen.17i18 Since TiIaZ- is not sufficiently stable in any of the nonaqueous solvents tried, only the spectrum of solid [ (C2Hj)qN]2Ti16 was obtained. Because of the very high molar extinction coefficients of the Ti&'- (in contrast to the much lower values for the electron-transfer spectra of the lanthanide MX63- c o m p l e ~ e s ~ ~ ~ ) , the spectrum of [ ( C ~ H ~ ) ~ N ] Z T ~ I ~

(16) S. K. Gupta, J . A m . Chem. Soc., 36, 747 (1914). (1T) C. K. Jrirgensen in "Halogen Chemistry," V. Gutman, Ed . , Academic

(18) C. K. Jplrgensen in "Chemical Applications of Spectroscopy," B. G. Press, London, 1967, pp 265-401.

Wybourne, Ed., Interscience Publishers, S e w York, N. Y., in press.

was difficult to obtain. In order to obtain good mull spectra of this highly absorbing substance, very thiii mulls with very fine particle sizes were required. Un- like the case of the 41114- salt discussed later, the instability of the compound made preparation of such mulls difficult. As a result of this and the uncertainty in the base line due to possible mismatch of light scatter of sample and reference, the accuracy of the positions of the transitions and to an even greater degree the relative intensities for the transitions (particularly the shoul- ders) of [(C*Hj)*N]2TiI6 in Table I is not so great as those for TiC16?- and TiBra2-. The TiCIe2- and TiBrs2- data are for acetonitrile solutions containing excess of the respective halide along with a small amount of the respective hydrogen halide to dehy- dratel9 the solutions and prevent hydrolysis.

The values in Table I are listed in such a way that the values for TiBra2- occur directly below those for the corresponding transitions for TiCle2- (the TiC1,j2- and TiBr6*- spectra are almost identical in appearance except for energy shift). The data for [(CZHZ)~N]Z- Ti16 are listed under what n70uld appear to be the corresponding transitions for TiBr6*- and TiC1e2- based on the energy differences expected due to differ- ences in electronegativities of the halides. Since thc 25.0-kK shoulder for Tic&'- and the 17.7-kK shoulder for TiBrsZ- are, relatively, too weak to be observed with certainty in a mull spectrum, it may be that a similar very weak shoulder might have been missed for [(C*HJ4- N12Ti16 and the 9.5-kK shoulder observed for this solid might correspond instead to the 29.6- and 21.7-kK shoulders for TiCle2- and TiBrs", respectively. The data of Brisdon, et a1.,6 indicate a considerable difference in the relative intensities of the transitions between the solution and solid spectra of and TiBrs*- and even an appreciable difference between the relative intensities of the transitions between two different salts of the same TiX6*-. Because of this an exact com- parison is not possible and i t is probable that the transi- tions match as shown in Table I. Considering such difference in relative intensities of transitions between ions in solutions and in the solids, the data for [(C*H,)4- NIzTiI6 in Table I are quite consistent with what is expected for the TiIe2- ion.

As discussed in the preceding section, the FeL- ion is considerably more stable than the other iodo com- plexes discussed here, and the absorption spectrum of this ion was obtained in both acetone and nitromethane in which there was no difference. The absorption spectrum (although more difficult to obtain because of

(19) J. L. Ryan in "Progress in Coordination Chemistry" (Proceedings of the 11th International Conference on Coordination Chemistry), M. Cais, Ed., Elsevier Publishing Co., N e w York, N. Y., 1968, paper D16, p 220.

VoZ. 8, No. 10, October 1969 WEAK OR UNSTABLE IODO COMPLEXES 2061

the shorter life of the FeIk-) also appears to be identical in all respects when a large excess of I- as (C4H,)4NI and a small amount of anhydrous H I is present in the acetone solution. This indicates that the FeI4- com- plex does not dissociate in dry acetone or nitromethane. FeC14- and FeBrd- both show relatively weak and narrow internal d transitions a t energies just below the start of the intense electron-transfer bands. At low temperature nine of these transitions are observed for FeCIc and eight are observed for FeBra- with the bands for FeBrl- a t 2 kK lower energy than for the chloride.z0 The complete assignment of these bands is not firmly established since more bands are observed than are expected for t d symmetry.17 Three similar bands are observed for Fe14- a t 9.615, 10.300, and 11.000 kK having molar extinction coefficients approxi- mately 60, 80, and 200, respectively, with others pre- sumably hidden under the first electron-transfer band. This gives an average shift for the first three bands of 2.3 kK to lower energy for Fe14- vs. FeBr4- due to in- creased covalency. It is interesting to note that the molar extinction coefficients of these transitions in- crease markedly with increase in halide mass. About a 30-fold average increase is observedz1 in going from FeCl1- to FeBr,-, and the transitions for Fe14- are about 5-10 times as intense as the corresponding transi- tions for FeBrl-. Assuming that these are spin-for- bidden bands (as suggested by Jgrgensen”), the in- creased covalency in going from the chloride to the iodide might be expected to cause relaxation of the forbiddenness and increase the observed intensities.

Day and JflrgensenlO have measured the electron- transfer spectra of several 3d group tetrahalides in- cluding FeC14- and FeBrl- and have discussed these in some detail. The electron-transfer spectra of FeX4- are compared in Table I1 where the results for FeC14- and FeBr4- are from ref 10. The trend toward lower molar extinction coefficients and increased structure with increase in halide mass observed by Day and Jflrgensen’O apparently continues with FeI4-. The change in energy of the electron-transfer bands in going from FeBr4- to Fe14- agrees very well with what would be expected from the difference between FeC14- and FeBr4- and the difference in electronegativities of the halides.

TABLE I1 ELECTRON-TRANSFER SPECTRA OF FeX4- a

FeC14- FeBr4- 21.20 [5800], (23.6), 25.50 [5960], (31.5), 35.65

Fell- 14 .3 148001, 17 .1 [5100], 19.0 [5300], 21.2 [6400],

a Wave numbers in kilokaisers with shoulders in parentheses FeI4- is in ace-

27.48 [7350], 31.8 [7600], (36. 8) , 41.2 [11 ,9001

[ 10,500]

24.6 [9900]

and molar extinction coefficients in brackets. tone.

The absorption spectra of the solid (CzHE)4N+ salts of AuC14-, AuBr4-, and AuI4- are shown in Figure 1.

(20) A. P. Ginsberg and M. B. Robin, IXOVE. Chem., 2, 817 (1963). (21) N. S. Gill, J. Chem. Soc., 3512 (1961).

\ L-------- I

I I I I I I I I , I 200 400 600 800 1000 1200

Wavelength, mp.

Figure 1.-Absorption spectra of solid gold( 111) tetrahalides: (1) (CoH&NAuCl4; (2) (CnH&NAuBr4; ( 3 ) ( CZHe)4iVAu14. Spectra 2 and 3 are displaced vertically.

The absorption spectrum of solid ( C ~ H ~ ) ~ N A U I ~ is very similar to those of the chloro and bromo salts with energy shifts of the order of magnitude expected from change of electronegativity of the halide. It is in- teresting to note the rather marked difference in appear- ance of the solution spectral2 and solid spectra of the chloro and bromo complexes (also observed for CsAu- C1422). It appears that this difference is due almost entirely to differences in relative intensities as indicated in Table I11 where the spectra of solid salts are com- pared with those of their solutions in acetonitrile. The acetonitrile spectra are essentially the same as those reported by Gangopadhayay and Chakrovorty12 except for the 18.2-kk shoulder for AuBr4- which was appar- ently overlooked by them. Table I11 is arranged in such a way that the transitions in a given column correspond most closely in shape and appearance in the solid spec- tra. The AuCl4- and AuBr4- solution spectra corre- spond to each other similarly and are matched to the solid spectra closest in energy.

TABLE I11 ABSORPTION SPECTRA OF AuX4- Q

AuC14- (CHsCN soh) (25.7), 31.1, 43 .8 (CzHs)aNAuCla (25.0), 29.8, (31.3), 37 .0 ,43 .5 AuBr4- (CHaCNsoln) (18.2), (21,1), 25.3, 3 9 . 1 (CzH&NAuBr4 (17.7), 19.6, (23.0), 28.2, -38 (CZHE)~NAUI~ (10.8), 12.7, 17 .0 ,23 .2

5 Notation as in Table I.

In the solid spectra, relatively strong transitions are observed which correspond to weak shoulders or transi- tions not resolved a t all in the solution spectra. I t has been proposed that these shoulders might be crystal field transitions within the d shell.12 If this were the case, the apparent increase in intensity niight be ex- pected if distortion of the square-planar AuX4- in the

(22) S. Yarnada and R. Tsuchida, Bull. C h e w Soc. Japan , 29, 424 1956).

2062 L). 1. BUSTIN, J. E, EAKLICY, ANI) A. A. VLEHK

crystal lattice destroyed the center of inversion. Such an explanation is not reasonable since not nearly so large an energy shift with change of halide would be expected for internal d transitions (less than 2-kK shift between successive halides is observed for internal d transitions of isoelectronic PtXd2-). l7 Thus all of the transitions of Table I11 appear to be electron-transfer transitions as also concluded by Jgrgensen. l7

It should be noted that the weak shoulders in the spectra of AuCld- and AuBr4- in solution are not due to complexes other than AuXd-. The absorption spectra are the same in various solvents except for very small energy shifts. Addition of excess C1- (up to saturated (CZH;)&Cl) to an acetonitrile solution of (CzH5)4NAu- Cl4 causes no change in absorption spectrum. Harris and R e e ~ e ~ ~ found that addition of Br- to a nitrometh- ane solution of iluBr4- markedly lowers its absorption in the visible region and attributed this to formation of AuBre3-. A similar effect was observed by this author except that the reportedz3 shift in the rather broad

(23) C. M. Harris and I. H. Keece, Natuve, 182, 1665 (1958).

absorption peak from 394 to 3'79 m l was not observed. The nitromethane cutoff occurs in this region and any reference mismatch may cause an apparent shift in a broad peak. \Then the experiment is done in acetoni- trile, the spectrum can be measured to lower wave- length. In freshly prepared (CdH9)jNBr-saturated so- lution the 394-in~ peak is reduced to less than 10% o€ its value in the absence of excess Br-, and a strong band develops a t 273 mp. This band has exactly the same position and shape as Br3- in this solvent, and the molar extinction coefficient is that of Br3- assuming 1 mol of Br3- produced per mole of AuBr4- disappear- ing. This 273-ml peak decays slowly with time as does that of Br3- due to slow reaction with the solvent. The conclusion is that Au(1II) is reduced in such solu- tions in qualitative agreement with results obtained in aqueous s o l ~ t i o n s , ~ ~ ~ ~ ~ and there is no evidence for gold(1II) halide complexes beyond AuX4-.

(24) B. I. Peschevitski, V. P. Kazakov, and A. &!I. Erenburg, Russ. J .

(25) V. P. Kazakov and %I. V. Konovalova, ibid., 13, 231 (1968). I?zoug. Chem., 8 , 437 (1963).

CONTRIBUTION FROM THE DEPARTMENT O F CHEMISTRY, GEORGETOWN UNIVERSITY, mTASHINGTON, D. c., AXD

THE J. HEYROVSKY INSTITUTE OF POLAROGRAPHY, CZECHOSLOVAK ACADEMY OF SCIENCES, PRAGUE, CZECHOSLOVAKIA

Study of the Pentacyanonitrosylchromate Ion. 111. Aquation

BY D. I. BUSTIN, J. E. EAKLEY, AND A. A. V L ~ E K

Received February 19, 1968

Aquation of Cr(CN)&03- in mildly acidic solutions yields Cr(CN)z(He0)3N0 which is inert to further aquation. tions of pH less than 2, further loss of CK- occurs. The slow step in the initial aquation is the loss of the first C S - . k = k'K[H+]/(l + K[H+]) , where K is 9 X lo2 M-' and k' is 7 X kcaljrnol.

In solu- The Cr-KO grouping remains intact until loss of CN- is complete.

This step is first order in substrate with a rate constant sec-' a t 17.5" in 1 M XaCl. Fork', AH* is 19

Several studies of the pentacyanonitrosylchromate ion, Cr(CN)5N03-, have led to the conclusion that this ion may be considered to contain Cr(I).I Although the ion is stable in basic or neutral solutions, it undergoes aquation in acidic solution. The ultimate products of this process are familiar Cr(II1) species. We have studied the kinetics of the multistep aquation and the nature of intermediates in order to observe the chemical properties of the Cr-NO grouping.

Experimental Section Reagents and Instr~ments.-K~Cr(CK)~NO was prepared by

rcduction of K2Cr04 by "SOH in the presence of KCNL and recrystallized three times from water. i\'aC104 was prepared by neutralization of HClO4, water was triply distilled from quartz, and other reagents were reagent grade used without further purifi- cation.

Spectrophotometric measurements mere made using a Cary 14 spectrophotometer and 1- and 5-cm cells. A Sargent XV polaro- graph and a V-301-Zbrojovka, Brno, polarograph (sensitivity of galvanometer used 3.33 X A/mm m ) were used for

(1) W. P. Griffith, J. Lewis, and G. Wilkinson, J . C ~ C ~ N . SOL., X i 2 (1950).

polarographic studies. The capillary employed had a drop time of 3.4 sec and flow rate of 2.05 mg/sec (in short circuit with sce and 64-cm Hg column). h modified Kalousek polarographic cell was used for measurements. No maximum suppressor was used. Ionic strength for polarographic measurements was generally maintained a t 1 M. Potentials were measured and are reported w . the sce .

A Metrohm E-336X potentiograph and Sargent combination electrode were used for pH titrations. A Sargent pH stat with combination electrode was used for kinetic measurements.

Results Kinetics of Primary Aquation.-Solutions of Cr-

(CN)5NO3--, containing NaCl as inert electrolyte, were allowed to undergo aquation a t various temperatures and at hydrogen ion concentrations controlled by the pH stat. In the course of aquation, 3.00 =+= 0.05 hy- drogen ions were absorbed per ion. Below pH 2.0 a slower, further absorption of hydrogen ions occurred. Figure 1 shows a logarithmic plot of a typical run. There is definite deviation from linearity at the be- ginning of the reaction. Similar deviations during the first third of the reaction were present in all runs.